[Openicc] colord 0.1.0 released!

[edmund ronald]

At the risk of being snarky
1. I think there are at most 100 people in the world who can
understand this discussion, combining the knowledge of color
management and that of the Mac print system.
2. Most developers don't have this knowledge.  I don't.
3. Most apps developed in this over-complicated system will be totally buggy
4 I spent weeks of my time reverse engineering the *effects* of the
Mac printing system with simple files
5 Why the hell do we accept this over-complicated model to work with?

Edmund

On Thu, Jan 20, 2011 at 7:05 AM, Chris Murphy <lists at colorremedies.com> wrote:
>
>
> On Jan 19, 2011, at 10:09 PM, Leonard Rosenthol wrote:
>
> On Wed, Jan 19, 2011 at 10:32 PM, Chris Murphy <lists at colorremedies.com>
> wrote:
>>
>> ICC spec (7.2.5) identifies 'prtr' as an "Output Device" profile.  So if
>> you really do an RGB-based output device, then you could have a 'prtr'
>> profile for it - but I am not aware of any in actual use.
>>
>> I'm really confused. Every single desktop inkjet printer comes with such
>> profiles.
>
> Not as output profiles they don't.  They may have proofing/simulation
> profiles in RGB - but not an actual prtr.  if you've got one, I'd love to
> see it...
>
> These are not proofing/simulation profiles. They are output device class
> profiles. I have hundreds of them. ColorSync Utility reports them as:
> Class: Output
> Space: RGB
> PCS: Lab
> And iccToXML reports them as:
>     <ProfileDeviceClass>prtr</ProfileDeviceClass>
>     <DataColourSpace>RGB </DataColourSpace>
> This is the same class as a GRACoL or SWOP press profile: prtr. The
> difference is just the colorspace is CMYK instead.
>
>
>
>>
>> But that's exactly what's being used by Photoshop and Lightroom,
>
> As a SOURCE profile but not as a PRINTER profile - and not at print time.
>
> No, if I'm a photographer printing my image from Lightroom the source
> profile is ProPhoto RGB which is class mntr, and the destination profile
> (the printer profile) I choose is
> EpsonStylusPro3880_3885_3890EpsonProofingPaperWhiteSemimatte which is class
> prtr and is RGB. It's an RGB to RGB conversion by the application, and it
> asks the OS to write out that prematched/preconverted data to a PDF print
> spool file and then tag it TWICE:
> The image object is
> tagged EpsonStylusPro3880_3885_3890EpsonProofingPaperWhiteSemimatte.icc.
> The OutputIntent is set
> to EpsonStylusPro3880_3885_3890EpsonProofingPaperWhiteSemimatte.icc.
> Same prtr class RGB profile.
>
>
>
>>
>> and dropped into the Mac OS PDF print spool file as an OutputIntent. It's
>> an RGB-based OutputIntent. All of these desktop printers are treated as RGB
>> 100% of the time when driven by the manufacturer's print driver.
>
> Not as OutputIntent.   And even if it was, the OI is only used in a PDF/X
> file.  It's ignored (as per spec) for any standard PDF.
> Perhaps you mean as a source profile - because that's what Quartz does.
>
> It can does do both. It just doesn't ensure both the object tag (the source
> space) and the OutputIntent (destination space) are the same.
> And regardless of whether the OutputIntent should be ignored, since the PDF
> isn't actually PDF/X, we are talking about Apple. They couldn't make it
> valid PDF/X and use version 4 ICC profiles anyway since they aren't allowed
> at least not in X-3.
>
> Chris
> _______________________________________________
> openicc mailing list
> openicc at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/openicc
>